This paper presents a multi-cue based PSO guided PF tracking
framework. there are two stages improved by employing PSO
algorithm in the particle filter framework, the first one is state
transition stage of particle filter framework, PSO algorithm is
utilized to make PF-particles move toward high likelihood area,another one is update stage of particle filter framework, PSO algorithm is used to search the optimal multi-cue weight configuration. Finally, the posterior state can be estimated by the relocated PF-particles. Experiments performed on three challenging public infrared video sequences demonstrate that our proposed tracking framework achieves considerable performance.